blockstorage

Block storage is a data storage method in which data is stored in fixed-size blocks, each with its own address. Unlike file or object storage, a block device presents raw storage to the operating system, which formats it with a filesystem or uses it directly by applications. Block storage is typically provided as volumes that are carved from a storage array, a direct-attached drive, or a cloud storage pool, and is commonly used for performance-sensitive workloads.

In practice, a host attaches a block device and the OS sees a block device (for example /dev/sdX). The storage backend may use SAN technologies such as Fibre Channel or iSCSI in traditional deployments, or NVMe over Fabrics for faster access. In cloud environments, services like persistent disks or block storage volumes are attached to virtual machines. Block storage is optimized for low latency and random I/O rather than large sequential files.

Key features include snapshots, cloning, thin provisioning, and replication; data can be encrypted at rest and in transit. Durability and availability depend on the underlying hardware or service's replication, RAID, erasure coding, and backups. Performance is influenced by IOPS, throughput, latency, and the interconnect; many systems support tiering and caching to balance cost and performance.

Use cases include databases, transactional systems, virtual machine disks, and other I/O-intensive applications. It is typically contrasted with file storage, which presents a hierarchical filesystem, and object storage, which stores data as discrete objects with metadata. Administrators manage capacity planning, provisioning models (thick vs thin), and data protection strategies.